AMD EPYC vs. Intel Xeon vs. POWER9 Linux Benchmarks
AMD EPYC vs. Intel Xeon open-source server benchmarks on Ubuntu Linux. Latest as of November 2018. Tests by Michael Larabel for a future article.
HTML result view exported from: https://openbenchmarking.org/result/1811168-SK-EPYCPOWER73&grr&rdt.
MariaDB
Clients: 64
Numpy Benchmark
Blender
Blend File: Barbershop - Compute: CPU-Only
Open Porous Media Git
OPM Benchmark: Flow MPI Extra - Threads: 16 - Solver Time
Blender
Blend File: Classroom - Compute: CPU-Only
VP9 libvpx Encoding
vpxenc
Open Porous Media Git
OPM Benchmark: Flow MPI Extra - Threads: 16
Blender
Blend File: Pabellon Barcelona - Compute: CPU-Only
FFTW
Build: Float + SSE - Size: 2D FFT Size 4096
Timed GCC Compilation
Time To Compile
Radiance Benchmark
Test: Serial
Blender
Blend File: Barbershop - Compute: CPU-Only
Open Porous Media Git
OPM Benchmark: Flow MPI Extra - Threads: 16 - Solver Time
Blender
Blend File: Fishy Cat - Compute: CPU-Only
Open Porous Media Git
OPM Benchmark: Flow MPI Extra - Threads: 32 - Solver Time
Open Porous Media Git
OPM Benchmark: Flow MPI Extra - Threads: 32
Open Porous Media Git
OPM Benchmark: Flow MPI Extra - Threads: 32 - Solver Time
FFTW
Build: Stock - Size: 2D FFT Size 4096
Tinymembench
Standard Memset
Tinymembench
Standard Memcpy
RAR Compression
Linux Source Tree Archiving To RAR
Blender
Blend File: BMW27 - Compute: CPU-Only
asmFish
1024 Hash Memory, 26 Depth
Fhourstones
Complex Connect-4 Solving
Blender
Blend File: Pabellon Barcelona - Compute: CPU-Only
Hierarchical INTegration
Test: FLOAT
Minion
Benchmark: Quasigroup
PostgreSQL pgbench
Scaling: Buffer Test - Test: Single Thread - Mode: Read Only
GNU GMP GMPbench
Total Time
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 32 - Solver Time
PostgreSQL pgbench
Scaling: Buffer Test - Test: Normal Load - Mode: Read Only
FFTW
Build: Float + SSE - Size: 2D FFT Size 2048
CacheBench
Test: Write
CacheBench
Test: Read / Modify / Write
CacheBench
Test: Read
BYTE Unix Benchmark
Computational Test: Dhrystone 2
MBW
Test: Memory Copy, Fixed Block Size - Array Size: 8192 MiB
Blender
Blend File: Classroom - Compute: CPU-Only
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 32
Memcached mcperf
Method: Add
Timed LLVM Compilation
Time To Compile
Timed Linux Kernel Compilation
Time To Compile
Radiance Benchmark
Test: SMP Parallel
Stockfish
Total Time
MBW
Test: Memory Copy - Array Size: 8192 MiB
7-Zip Compression
Compress Speed Test
NAMD
ATPase Simulation - 327,506 Atoms
Minion
Benchmark: Solitaire
Memcached mcperf
Method: Prepend
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 32 - Solver Time
Blender
Blend File: Fishy Cat - Compute: CPU-Only
Memcached mcperf
Method: Set
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 16 - Solver Time
GraphicsMagick
Operation: Noise-Gaussian
Timed PHP Compilation
Time To Compile
R Benchmark
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 16
IndigoBench
Scene: Supercar
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 16 - Solver Time
Stress-NG
Test: Context Switching
Memcached mcperf
Method: Replace
Memcached mcperf
Method: Append
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 64 - Solver Time
Open Porous Media Git
OPM Benchmark: Flow MPI Norne - Threads: 64
AOBench
Size: 2048 x 2048 - Total Time
IndigoBench
Scene: Bedroom
Memcached mcperf
Method: Get
MBW
Test: Memory Copy, Fixed Block Size - Array Size: 4096 MiB
Chaos Group V-RAY
Mode: CPU
Minion
Benchmark: Graceful
FFTW
Build: Stock - Size: 2D FFT Size 2048
GraphicsMagick
Operation: Sharpen
GraphicsMagick
Operation: Enhanced
GraphicsMagick
Operation: Swirl
GraphicsMagick
Operation: Rotate
GraphicsMagick
Operation: HWB Color Space
Rust Mandelbrot
Time To Complete Serial/Parallel Mandelbrot
Apache Benchmark
Static Web Page Serving
MBW
Test: Memory Copy - Array Size: 4096 MiB
Blender
Blend File: BMW27 - Compute: CPU-Only
m-queens
Time To Solve
Memcached mcperf
Method: Delete
GNU MPC
Multi-Precision Benchmark
NAS Parallel Benchmarks
Test / Class: LU.C
Stress-NG
Test: Socket Activity
C-Ray
Total Time - 4K, 16 Rays Per Pixel
lzbench
Test: XZ 0 - Process: Decompression
lzbench
Test: XZ 0 - Process: Compression
x265
H.265 Video Encoding
Gzip Compression
Linux Source Tree Archiving To .tar.gz
John The Ripper
Test: Traditional DES
ebizzy
Rodinia
Test: OpenMP Streamcluster
Timed ImageMagick Compilation
Time To Compile
Tesseract OCR
Time To OCR 7 Images
eSpeak Speech Engine
Text-To-Speech Synthesis
POV-Ray
Trace Time
Java SciMark
Computational Test: Composite
Sunflow Rendering System
Global Illumination + Image Synthesis
Compile Bench
Test: Compile
Cryptsetup
PBKDF2-whirlpool
Cryptsetup
PBKDF2-sha512
GNU Octave Benchmark
PyBench
Total For Average Test Times
lzbench
Test: Brotli 0 - Process: Decompression
lzbench
Test: Brotli 0 - Process: Compression
Cryptsetup
PBKDF2-whirlpool
lzbench
Test: Libdeflate 1 - Process: Decompression
lzbench
Test: Libdeflate 1 - Process: Compression
Crafty
Elapsed Time
lzbench
Test: Zstd 1 - Process: Decompression
lzbench
Test: Zstd 1 - Process: Compression
Node.js Octane Benchmark
FFTW
Build: Float + SSE - Size: 2D FFT Size 1024
Y-Cruncher
Calculating 500M Pi Digits
Zstd Compression
Compressing ubuntu-16.04.3-server-i386.img, Compression Level 19
Timed Apache Compilation
Time To Compile
Stress-NG
Test: CPU Stress
Aircrack-ng
Stress-NG
Test: Matrix Math
Stress-NG
Test: Crypto
Stress-NG
Test: Vector Math
Stream
Type: Copy
Swet
Average
SciMark
Computational Test: Composite
Mencoder
AVI To LAVC
John The Ripper
Test: Blowfish
LuaJIT
Test: Composite
FLAC Audio Encoding
WAV To FLAC
Rodinia
Test: OpenMP LavaMD
GIMP
Test: auto-levels
PHPBench
PHP Benchmark Suite
Go Benchmarks
Test: build
Redis
Test: SADD
RAMspeed SMP
Type: Copy - Benchmark: Integer
RAMspeed SMP
Type: Add - Benchmark: Integer
RAMspeed SMP
Type: Average - Benchmark: Integer
Rust Prime Benchmark
Prime Number Test To 200,000,000
RAMspeed SMP
Type: Scale - Benchmark: Floating Point
RAMspeed SMP
Type: Average - Benchmark: Floating Point
GIMP
Test: rotate
RAMspeed SMP
Type: Triad - Benchmark: Floating Point
RAMspeed SMP
Type: Add - Benchmark: Floating Point
RAMspeed SMP
Type: Copy - Benchmark: Floating Point
Primesieve
1e12 Prime Number Generation
OpenSSL
RSA 4096-bit Performance
GnuPG
Linux 4.3 Package File Encryption
Redis
Test: SET
Sudokut
Total Time
dcraw
RAW To PPM Image Conversion
NAS Parallel Benchmarks
Test / Class: EP.C
MBW
Test: Memory Copy, Fixed Block Size - Array Size: 1024 MiB
x264
H.264 Video Encoding
Redis
Test: LPUSH
FFTW
Build: Stock - Size: 2D FFT Size 1024
GIMP
Test: resize
Redis
Test: GET
Bork File Encrypter
File Encryption Time
FFTW
Build: Float + SSE - Size: 1D FFT Size 4096
FFmpeg
H.264 HD To NTSC DV
DaCapo Benchmark
Java Test: Jython
LAME MP3 Encoding
WAV To MP3
MBW
Test: Memory Copy - Array Size: 1024 MiB
Timed MAFFT Alignment
Multiple Sequence Alignment
N-Queens
Elapsed Time
FFTW
Build: Float + SSE - Size: 1D FFT Size 2048
Sysbench
Test: CPU
Timed HMMer Search
Pfam Database Search
FFTW
Build: Float + SSE - Size: 2D FFT Size 512
Darktable
Test: Masskrug - Acceleration: CPU-only
MBW
Test: Memory Copy, Fixed Block Size - Array Size: 512 MiB
Ogg Encoding
WAV To Ogg
Smallpt
Global Illumination Renderer; 128 Samples
Git
Time To Complete Common Git Commands
FFTW
Build: Float + SSE - Size: 1D FFT Size 1024
OSBench
Test: Memory Allocations
FFTW
Build: Float + SSE - Size: 1D FFT Size 32
FFTW
Build: Stock - Size: 1D FFT Size 4096
libjpeg-turbo tjbench
Test: Decompression Throughput
PolyBench-C
Test: Correlation Computation
PolyBench-C
Test: Covariance Computation
Bullet Physics Engine
Test: Raytests
FFTW
Build: Float + SSE - Size: 1D FFT Size 256
OSBench
Test: Create Files
NAS Parallel Benchmarks
Test / Class: LU.A
MBW
Test: Memory Copy - Array Size: 512 MiB
FFTW
Build: Float + SSE - Size: 2D FFT Size 256
Optcarrot
Optimized Benchmark
FFTW
Build: Float + SSE - Size: 1D FFT Size 512
FFTW
Build: Stock - Size: 2D FFT Size 512
FFTW
Build: Stock - Size: 1D FFT Size 2048
Tachyon
Total Time
FFTW
Build: Float + SSE - Size: 1D FFT Size 64
FFTW
Build: Float + SSE - Size: 2D FFT Size 32
FFTW
Build: Stock - Size: 2D FFT Size 256
Parboil
Test: OpenMP CUTCP
PolyBench-C
Test: 3 Matrix Multiplications
FFTW
Build: Stock - Size: 1D FFT Size 1024
FFTW
Build: Stock - Size: 2D FFT Size 32
FFTW
Build: Float + SSE - Size: 1D FFT Size 128
FFTW
Build: Stock - Size: 2D FFT Size 64
FFTW
Build: Float + SSE - Size: 2D FFT Size 128
FFTW
Build: Float + SSE - Size: 2D FFT Size 64
FFTW
Build: Stock - Size: 1D FFT Size 64
FFTW
Build: Stock - Size: 1D FFT Size 512
FFTW
Build: Stock - Size: 2D FFT Size 128
TSCP
AI Chess Performance
FFTW
Build: Stock - Size: 1D FFT Size 128
FFTW
Build: Stock - Size: 1D FFT Size 256
FFTW
Build: Stock - Size: 1D FFT Size 32
glibc bench
Benchmark: sin
glibc bench
Benchmark: cos
MBW
Test: Memory Copy, Fixed Block Size - Array Size: 128 MiB
glibc bench
Benchmark: tanh
glibc bench
Benchmark: pthread_once
glibc bench
Benchmark: ffs
glibc bench
Benchmark: ffsll
glibc bench
Benchmark: sqrt
MBW
Test: Memory Copy - Array Size: 128 MiB
Java SciMark
Computational Test: Jacobi Successive Over-Relaxation
Java SciMark
Computational Test: Dense LU Matrix Factorization
Java SciMark
Computational Test: Sparse Matrix Multiply
Java SciMark
Computational Test: Fast Fourier Transform
Java SciMark
Computational Test: Monte Carlo
Bullet Physics Engine
Test: Convex Trimesh
Bullet Physics Engine
Test: Prim Trimesh
Bullet Physics Engine
Test: 136 Ragdolls
Bullet Physics Engine
Test: 1000 Convex
Bullet Physics Engine
Test: 1000 Stack
Bullet Physics Engine
Test: 3000 Fall
SciMark
Computational Test: Jacobi Successive Over-Relaxation
SciMark
Computational Test: Dense LU Matrix Factorization
SciMark
Computational Test: Sparse Matrix Multiply
SciMark
Computational Test: Fast Fourier Transform
SciMark
Computational Test: Monte Carlo
LuaJIT
Test: Jacobi Successive Over-Relaxation
LuaJIT
Test: Dense LU Matrix Factorization
LuaJIT
Test: Sparse Matrix Multiply
LuaJIT
Test: Fast Fourier Transform
LuaJIT
Test: Monte Carlo
Stream
Type: Add
Stream
Type: Triad
Stream
Type: Scale
Compile Bench
Test: Read Compiled Tree
Phoronix Test Suite v10.8.4